Test 8 failures. Trying to diagnose CPU, Mobo, or ram.

[Hybrid Divide]

The high prices of RAM had me initiate my next PC build about a year earlier than I'd have liked.

Had to buy used RAM, unfortunately. So, of course before I even installed Windows, I updated the BIOS and ran Memtest86. (CPU and Motherboard are brand new)


Running at the Expo speed profile, everything seemed fine until it got to Test 8 (Random Number Sequence.) Got a bunch of errors.
Ran the test again at stock speeds and passed.
Ran Test 8 only with one stick at Expo speeds, then the other. Both passed.
Ran Test 8 only with both sticks. Passed.
Ran another full test with both. Failed at Test 8 again.
Ran a final Test 8 only with both. Failed.
Ran full tests on both sticks individually at Expo speeds. Passed.
Ran full tests on both sticks individually in slot B2 with expo speeds. Passed.
Ran a combined test both sticks with the Expo 2 profile. Failed.

Relevant specs:
Motherboard: ASUS ROG Strix X870E-E Gaming WiFi.
CPU: AMD Ryzen 7 9800X3D
RAM: G.Skill Trident Z Neo RGB DDR5 64GB (2x32GB 6000 mt/s)


I COULD run at the stock speeds, but I paid for 6000mt/s, I don't want to get less than that, and I worry that it could get worse in the future. I admit I don't really know much about tuning RAM aside from setting the Expo/XMP profile. Not new to PC builds, but never really did any overclocking.


I should still be able to return the RAM, but given the prices, I'll pay even MORE to get the same kind of RAM now as when I did a few weeks ago. So I'm really hoping it's the CPU or the motherboard that's the real problem. They'd be a pain to return, too. But dealing with Amazon is easier than ebay. Faster, too. And that's before we even get into the problem of RAM being even more expensive now.

Any advice on narrowing down the cause?


I have enclosed the log file from the first run. (Pretty sure I picked the right one!)

Any advice would be appreciated.

[David (PassMark)]

It isn't unusual to have RAM to fail in dual channel mode, but be OK in single channel mode. The access patterns & environment (voltages & EMI) can be different enough to provoke an error if the part was marginal.

There is a good chance that the problem is the RAM and not the CPU nor motherboard.

There is also some chance that the 2nd hand RAM was being sold because they knew, or suspected, it was bad.

Maybe you can tweak the voltages and timings to get it stable, but you shouldn't need to, and it is very hard & slow to know if you really have achieved stability (or just slightly lowered the probability of errors). Also easy to make things worse. May as well just use the slower default JEDEC profile if you are going to do that.

[Hybrid Divide]

It isn't unusual to have RAM to fail in dual channel mode, but be OK in single channel mode. The access patterns & environment (voltages & EMI) can be different enough to provoke an error if the part was marginal.

There is a good chance that the problem is the RAM and not the CPU nor motherboard.

There is also some chance that the 2nd hand RAM was being sold because they knew, or suspected, it was bad.

Maybe you can tweak the voltages and timings to get it stable, but you shouldn't need to, and it is very hard & slow to know if you really have achieved stability (or just slightly lowered the probability of errors). Also easy to make things worse. May as well just use the slower default JEDEC profile if you are going to do that.

Thanks for the reply. While I don't have proof that it was working for the guy who sold the ram, he has been helpful in providing suggestions, and insists the ram worked for him with expo speeds and passed a memtest. Again, I know that isn't proof, but my gut tells me the guy isn't lying.